SM

Aa

8b

8

a

ARRA

KCSSTTU

I

iitBeaeais2di

h1

Urban Forestry & Urban Greening 14 (2015) 1110–1121

Contents lists available at ScienceDirect

Urban Forestry & Urban Greening

j ourna l h om epage: www.elsev ier .com/ locate /u fug

tructure and ecosystem services of small-leaved lime (Tilia cordataill.) and black locust (Robinia pseudoacacia L.) in urban environments

. Mosera,∗, T. Rötzera, S. Pauleitb, H. Pretzscha

Chair for Forest Growth and Yield Science, School of Life Sciences Weihenstephan, Technische Universität München, Hans-Carl-von-Carlowitz-Platz 2,5354 Freising, GermanyChair for Strategic Landscape Planning and Management, School of Life Sciences Weihenstephan, Technische Universität München, Emil-Ramann-Str. 6,5354 Freising, Germany

r t i c l e i n f o

rticle history:eceived 13 March 2015eceived in revised form 6 October 2015ccepted 11 October 2015

eywords:arbon storagehadingite conditionranspirationree allometry

a b s t r a c t

Knowledge of the structure and morphology of common urban trees is scarce, particularly of the full lifecycle of a tree. The present and future structural dimensions of urban tree species are of an increasinginterest because tree growth is associated with its ecosystem services. The purpose of this study was tocharacterize the dimensions of two urban tree species (small-leaved lime, Tilia cordata Mill. and blacklocust, Robinia pseudoacacia L.) and to predict future structural dimensions based on the diameter atbreast height and tree age. Regression equations were developed for tree height, crown diameter, crownheight, crown volume, crown projection area, and open surface area of the tree pit. The data revealedstrong relationships (r2 > 0.7) between crown diameter, crown volume, crown projection area, crownheight, tree pit for both species, and tree height of T. cordata. Based on tree dimensions and the leaf areaindex (LAI), three ecosystem services (carbon storage, shading, and cooling effects) were estimated for

rban trees the analyzed trees. The results indicated that urban trees considerably improved the climate in cities,with carbon storage, shading, and cooling of single trees showing a direct relationship with LAI and age.The associations of tree growth patterns identified in this study can be used as guidelines for tree plantingin cities and their ecosystem services; they may improve the management and planning of urban greenareas.

© 2015 Elsevier GmbH. All rights reserved.

ntroduction

Urban trees are a central element in green space planning. Dur-ng tree planting, city planners need to consider the fitness of treesn their specific environment and their future growth. However,he establishment of trees in urban areas faces several restrictions.ecause of various influences, including limited root volume (Dayt al., 1995; Grabosky and Bassuk, 1995), soil compaction (Beattynd Heckman, 1981; Day et al., 1995), high temperatures (Akbarit al., 2001; Kjelgren and Clark, 1992), less water availability (Beattynd Heckman, 1981; Whitlow and Bassuk, 1986), and mechanicalnjury (Beatty and Heckman, 1981; Foster and Blaine, 1978), urbanites considerably differ from forests (Nowak et al., 1990; Quigley,

004). These stressful conditions in urban environments can hin-er tree development. The growth and the corresponding tree age

n urban areas, particularly of newly planted trees along streets,

∗ Corresponding author. Tel.: +49 8161 71 4719.E-mail address: astrid.moser@lrz.tum.de (A. Moser).

ttp://dx.doi.org/10.1016/j.ufug.2015.10.005618-8667/© 2015 Elsevier GmbH. All rights reserved.

are often more limited than those of trees in parks or rural areas(Foster and Blaine, 1978; Rhoades and Stipes, 1999). Therefore,knowledge of the growth dimensions of tree species in associationwith the stand conditions is important. For selecting the right treespecies for a location, quantitative information on tree growth fromplanting to maturity and decline is required (Peper et al., 2014).However, knowledge of the size and growth at a particular age andthe effects of urban climates on tree growth remain unclear andless researched (Kjelgren and Clark, 1992; Peper et al., 2001a; Rust,2014).

Modelling of tree growth is usually based on diameter at breastheight (dbh). According to the pipe model theory (Chiba, 1998;Shinozaki et al., 1964a,b) and the functional carbon balance theory(Mäkelä, 1990), tree structures (e.g., crown volume) can be esti-mated from tree dimensions (e.g., dbh). Several studies are basedon these theories and use dbh as an explanatory variable for for-est stands, predicting tree height and crown dimensions (Pretzsch

et al., 2012; Stage, 1973; Watt and Kirschbaum, 2011). McMahon(1975), however, promoted three similarity models to describegrowth of trees as power law functions (Dahle and Grabosky, 2009).Structural developments of urban trees are of interest because of

Urban

afpa11(Ha2N

(owCysrlbPbsa1tP

ppum(es2tat(fu(vbceaas

-

-

-

-

A. Moser et al. / Urban Forestry &

direct association between tree dimensions and their ecologicaleatures and benefits (Troxel et al., 2013). Moreover, urban treesrovide numerous ecosystem services to moderate microclimatesnd improve environmental conditions (Bolund and Hunhammar,999; Troxel et al., 2013), including carbon storage (McPherson,998; Nowak and Crane, 2002), mitigation of the heat island effectAkbari et al., 2001), reduction of rainwater runoff (Bolund andunhammar, 1999; Xiao et al., 2000a,b), pollutant filtering (Bolundnd Hunhammar, 1999; McPherson et al., 1997; Pretzsch et al.,015), shading (Akbari et al., 2001), and cooling (Dimoudi andikolopoulou, 2003; Shashua-Bar and Hoffman, 2003).

Tree parameters are associated with these ecosystem servicesFig. 1). For example carbon storage-closely linked to the biomassf the tree-increases steadily with age (Var. 2, Yoon et al. (2013)),hile the shading intensity of trees decreases with age (Var. 1,hreptun 2015), whereas the leaf area index of a stand peaks at aoung tree age (Var. 3, Ryan et al., 1997; Nock et al., 2008). Severaltudies (Scott et al., 1998; Stoffberg et al., 2008; Xiao et al., 2000a,b)evealed a direct association of these benefits with tree canopy andeaf area. The crown projection area and crown volume form theasis for leaf area calculation (Binkley et al., 2013; Pretzsch, 2014;retzsch et al., 2015). Precise leaf area estimates are importantecause leaf area influences key physiological processes and atmo-pheric fluxes such as air pollutant and rainfall interception (Pepernd McPherson, 1998), photosynthesis (Kramer and Kozlowski,979), evapotranspiration (McPherson and Peper, 2012), respira-ion (McPherson and Peper, 2012), and shading (McPherson andeper, 2012; Rijal et al., 2012).

Methods for predicting tree dimensions, including the crownrojection area, crown volume, and leaf area, enable arborists,lanners, and researchers to model the growth and benefits ofrban trees. Based on tree allometry, alternative and improvedanagement practices for urban trees can be defined and applied

McPherson et al., 2000; Peper et al., 2001b). However, most growthquations for urban trees have been developed for only a fewpecies with limited size and age ranges (Rust, 2014; Troxel et al.,013). Particularly, studies dealing with the development of urbanrees over time and changes in their ecosystem services in associ-tion with their growth are limited. The growth of various urbanrees in US cities was studied by Peper et al. (2001a,b), Peper et al.2014), Rijal et al. (2012), and Troxel et al. (2013). However, only aew common European species have been analyzed to date, partic-larly regarding their ecological features. Larsen and Kristoffersen2002) and Hasenauer (1997) studied the growth of Tilia cordata, aery common urban tree species of temperate regions. Althoughoth studies provided information on the growth patterns of T.ordata over time, they did not establish associations with theircosystem services. To the best of our knowledge, the dimensionsnd growth patterns of other urban species have not been system-tically researched in European cities. Therefore, the aims of thistudy are

to investigate growth associations of two common urban treespecies covering several ecological niches: a shade-tolerant(small-leaved lime, T. cordata Mill.) and light-demanding (blacklocust, Robinia pseudoacacia L.) species,

to establish tree growth associations from young, newly plantedto very old trees,

to predict tree height and crown dimensions (diameter, height,volume, and projection area) on the basis of stem diameter and

age for T. cordata and R. pseudoacacia trees in street canyons,public places, and parks from low to high tree age,

to determine and calculate ecosystem services of urban trees (car-bon storage, shading, and cooling) depending on their growth

Greening 14 (2015) 1110–1121 1111

dimensions and age in order to illustrate the impact of trees forurban climates,

- closing the knowledge gap about the relationships of tree param-eters with their ecosystem services and the change of theassociated services of urban trees with tree age.

Material and methods

Site description and data collection

Tree data were collected in München (48◦09′N, 11◦35′E, 519 ma.s.l.) and Würzburg (49◦48′N, 9◦56′E, 177 m a.s.l.), southernGermany. They considerably differ in their climatic characteris-tics. The long-term annual precipitation means of München andWürzburg are 959 mm and 596 mm, respectively, whereas themean annual temperature (1961–1990) in both cities is 9.1 ◦C(DWD, 2015). According to tree inventories, 750,000 trees havebeen planted in München and 40,000 in the municipal area ofWürzburg. For this study, 225 T. cordata trees and 195 R. pseudoa-cacia trees were selected for measurement. We chose these twospecies because they markedly differ regarding their ecological fea-tures and are two of the more common tree species in the two citiesand elsewhere (Pauleit et al., 2002). While T. cordata is a shade-tolerant species R. pseudoacacia requires a certain amount of light,therefore T. cordata is expected to have a higher leaf area index (LAI)than R. pseudoacacia. Tree selection was based on visual impres-sion, and damaged, pruned, or low-forking trees were excluded.Tree data was collected from November 2013 to September 2014.

All the measured trees were classified as either park trees,trees in public places, or street trees. Trees were considered parktrees when planted in a green area without buildings. Street treeswere trees planted in a street canyon. As trees in public places-smaller, mostly paved spaces freely accessible to the public-onlyfree-standing trees with open, detached crowns were selected. Foreach tree, the following information was recorded: diameter atbreast height (dbh), tree height, height to live crown base, crowndiameter, tree pit, vitality, coordinates and altitude, and distanceto adjacent buildings and trees.

The stem diameter of all trees was measured with a diametermeasurement tape at a height of 1.3 m. Because nursery workersand green space planners measure dbh at a height of 1 m, for the64 T. cordata and 84 R. pseudoacacia trees, stem diameters at 1-and 1.3-m tree height, respectively, were measured to calculatea conversion factor with linear regressions and related regressioncoefficients (a and b) of the form:

dbh (1.0 m) = a + b × dbh (1.3 m) (1)

Tree height and crown base height (from the lowest primarybranch to the top of the crown) were calculated using a VertexForestor. Crown radii and tree pit (open surface area) were mea-sured in eight intercardinal directions (N, NE, . . ., NW) along theground surface with a measuring tape from the center of the trunkto the tip of the most remote downward-projecting shoot and to thelast visible open, non-asphalted surface of the soil. The distances tothe neighboring trees and buildings were estimated. The vitality ofthe trees was rated based on the scale of Roloff (2001), ranging fromvery good (0) to very poor (3) conditions regarding the branchingstructures of the crown.

Calculations based on measured data

From the measured data, crown radius, crown diame-ter, crown volume, crown projection area, and biomass were

1112 A. Moser et al. / Urban Forestry & Urban Greening 14 (2015) 1110–1121

velop

cm

c

wdd

c

C

c

eoS

L

iPtrpwsRBL(aadbtc

l

a

l

beech Fagus sylvatica were related to LAI [m2/m2] by regression

Fig. 1. Growth of an urban tree over time and hypothetical de

alculated. The mean crown radius cr was defined as the quadraticean:

r =√(

r2N + r2

NE + · · · + r2NW

)8

(2)

ith rN as the widest measured crown extension in the northernirection, . . ., rNW the widest crown extension in the northwestirection.

The crown projection area CPA and the crown volume cv werealculated as

PA = cr2 × � (3)

v = CPA × crown height (4)

To obtain the aboveground woody biomass of T. cordata, thequations of Cihák et al. (2014) were applied. The woody biomassf R. pseudoacacia was computed with the functions of Clark andchroeder (1986).

eaf area index

LAI was derived from hemispherical photographs captured dur-ng the fully leafed phase (July–August) using a Nikon Coolpix5100 camera with fisheye lens and Mid-OMount. Between theime of measurement and the fully leafed phase, some trees wereemoved or severely pruned, and these trees were excluded fromhotography. The resulting hemispherical photos were analyzedith the programs WinSCANOPY (Régent Instruments Inc.). This

oftware allows the calculation of the LAI by several methods. For. pseudoacacia, LAI (Bonhom)-Lin was used, which is based on theonhomme and Chartier (1972) linear method. For T. cordata, theAI (2000)-Lin calculated with the Licor LAI2000 linear methodMiller, 1967; Welles and Norman, 1991) resulted in the most reli-ble values. We performed regression analyses with generalizeddditive models (GAM, package mgvc) and mixed smoothed pre-ictor variables to obtain specific models for calculating LAI on theasis of the measured tree dimensions. The best fitting model withhe lowest Akaike Information Criterion (AIC) was chosen, for T.ordata resulting in:

n (LAI) = a + s (b1) × ln (dbh) + s (b2) × ln (h) + s (b3) × ln (2 × cr)

(5)

nd for R. pseudoacacia resulting in:

n (LAI) = a + s (b1) × ln (h) + s (b2) × ln (cv) + s (b3) × ln (CPA) (6)

ment of associated ecosystem services (Var1, Var2, and Var3).

with a as intercept, s as smooth function for every slope variable(b) and h as tree height.

Age estimation

The age of all trees was estimated based on the measured treeparameters dbh and tree height. For T. cordata, the formula ofLukaszkiewicz and Kosmala (2008) was applied:

age = a + e((b+c×dbh)/(100+d×h)) (7)

with a = 264.073, b = 5.5834, c = 0.3397, d = 0.0026, dbh in cm, and hin m.

To obtain the age of R. pseudoacacia, we multiplied dbh with aspecies-dependent age factor of 0.996, which was computed by themeasurements of Dwyer (2009) for Gleditsia triacanthos.

Ecosystem services

To evaluate the benefits of urban trees for the city climate, weestimated three ecosystem services as examples “carbon storage,shading, and transpiration by cooling”. To calculate the carbon stor-age of a tree, the biomass was multiplied with 0.5 following IPCC(2003) and Yoon et al. (2013). Shading can be obtained dependingon crown height hc and crown radius cr for each hour of the day by

AS = 2 × cr × hc × f × cot (Y) (8)

with AS = shaded area in m2, f = correction factor for the crown form,Y = elevation angle of the sun which in turn is a function of latitude,day of the year and hour of the day (e.g. DVWK, 1996), assuminga perpendicularly standing object on a horizontal surface (Häckel,2012).

As shading coefficient we used the average shaded area of allsunlight hours of June 21st. For rough estimations of transpiration,and thus, the cooling effect of individual trees, simulation resultsof the process-based growth model BALANCE were used. Thismodel simulates the growth and water balance of individual treesdepending on their environment (Rötzer et al., 2010). Subsequently,the annual transpiration sums of the deciduous species European

analysis:

traa [mm] = −0.0168 × LAI2 + 0.4555 × LAI + 0.2844(

r2 = 0.65)

(9)

A. Moser et al. / Urban Forestry & Urban Greening 14 (2015) 1110–1121 1113

Table 1Regression parameters for predicting diameter at breast height (dbh) at 1 m from dbh at 1.3 m for T. cordata and R. pseudoacacia of the form: dbh (1.0 m) = a + b × dbh (1.3 m).Listed are the regression coefficients (a and b) with standard error (SE), coefficients of determination (r2), F and p values, and sample size (n). Dbh was measured and calculatedin centimeters.

Species n Coefficients F r2 p

a SE b SE

1.01.0

at

S

wpactwtpdephupl(

l

cewra(

C

w

TMi

Tilia cordata 64 0.37 0.35

Robinia pseudoacacia 84 0.17 0.22

To estimate the average annual energy withdrawn from thetmosphere for transpiration, the transpiration sum must be mul-iplied by the potential evaporation energy [2.34 MJ (kg H2O)−1].

tatistical analysis

The software packages R, version 3.0.3 (R Core Team, 2014)as used for statistical analysis. First, regression analyses wereerformed to determine the associations between stem diameternd age, tree height, crown height, crown base, crown diameter,rown volume, crown projection area, and open surface area ofhe tree pit. Analysis of variance (ANOVA) with Tukey’s HSD testas then performed to identify differences between the measured

ree dimensions in the abovementioned categories (park, publiclace, street). Two Sample t-tests were applied to check for speciesifferences regarding the ecosystem services. Log transformationnsured homogeneity of variances. Normal distribution of data wasartially given, but is not a necessary assumption for ANOVA withigh sample size (Stevens, 1999). Assumption of normality of resid-als was assessed with graphical displace. All regressions wereerformed using log transformation of the tree dimensions, fol-

owing Pretzsch et al. (2012), Stoffberg et al. (2008), and Peper et al.2001a):

n (y) = a + b × ln (x) (10)

With ordinary least squares (OLS) regression, the response (y) isalculated from the predictor (x). Because in application of the mod-ls, the growth of trees over time (y) will be estimated based on x,e selected OLS instead of reduced major axis or moving average

egression (Niklas, 1994). For the back-transformation of the log-rithm, the bias correction (CF) by Baskerville (1971) and Sprugel1983) was applied:

F = e

(RSE2/2

)(11)

ith RSE = residual standard error.

able 2easured and calculated tree dimensions: dbh, tree height, crown height, crown diamete

n 5 ascending age classes of 20 years.

Age n DBH [cm] Tree height [m] Crown hei

Min Mean Max Min Mean Max Min Mea

Tilia cordata<20 27 6.4 9.6 13.0 5.0 6.2 8.0 2.5 3.820–40 86 11.9 21.7 31.0 6.1 10.3 16.1 3.2 7.440–60 83 27.2 37.1 46.5 9.8 14.3 18.6 6.6 10.960–80 22 44.8 51.7 60.7 14.2 18.1 23.3 9.7 14.5>80 8 63.0 78.1 107.0 11.4 19.1 27.1 7.0 15.6

Robinia pseudoacacia<20 34 7.6 13.7 21 5.7 9.6 18 3.1 6.220–40 67 21.2 30.4 40 7.3 13.6 24 4.7 9.640–60 47 31.5 45.1 60 9.1 16.0 28 6.2 11.560–80 33 41.1 57.4 80 12 16.8 27 8.4 13.0>80 14 60.9 84.8 102 14.7 21.3 27 11 17.3

2 0.01 6269 0.99 <0.0011 0.01 21,740 0.99 <0.001

Results and discussion

Stem diameter conversion

A regression analysis revealed highly significant relationshipsbetween stem diameters at 1 m and 1.3 m for both species. Theregression results are presented in Table 1.

Tree dimensions and growth equations

All measured and calculated tree dimensions are presented inTable 2. The minimum, maximum, and mean values are given inascending age classes. The results indicate similar growth patternsof both species, and with increasing age, the tree dimensions alsoincreased. Roman and Scatena (2011) cited several authors statingthe average age of urban trees ranging from 13 years (Skiera andMoll, 1992), 15 years (Nowak et al., 2004) up to 73 years in Acersaccharinum (Richards, 1979). The average age of the trees in thisstudy was 41 years for T. cordata and 39 years for R. pseudoacacia.Up to an age of approximately 80 years, the annual growth washigh (crown diameter of R. pseudoacacia increased on average by15 cm annually for all trees older than 60 years). For both species,a wide range of growth dimensions was included in the samples.The dbh of T. cordata ranged from 6.4 to 107 cm and that of R. pseu-doacacia from 7.6 to 102 cm. R. pseudoacacia had the largest size atages greater than 80 years, reaching a maximum stem diameter of102 cm, tree height of 27 m, and crown diameter of 20.1 m. Overall,we found markedly higher growth rates for R. pseudoacacia thanfor T. cordata, resulting in larger trees at equal ages. For example, atages of up to 20 years, T. cordata had an average tree height of 6.2 mand crown diameter of 3.1 m, whereas R. pseudoacacia reached anaverage height of 9.6 m and crown diameter of 5 m.

Linear regression analyses were used to determine the associ-ations between dbh, tree height, and crown dimensions (Table 3).

The regressions were highly significant at alpha levels of 0.05. Thehigh r2 (>0.7) for most tree variables indicates strong relationships.For T. cordata, dbh associated with the crown projection area, crownvolume, and crown diameter displayed the strongest dependences.

r, crown projection area (cpa), and crown volume of T. cordata and R. pseudoacacia

ght [m] Crown diameter [m] CPA [m2] Crown volume [m3]

n Max Min Mean Max Min Mean Max Min Mean Max

5.4 2.0 3.1 4.7 3 8 18 10 32 93 12.2 2.9 6.2 10.1 7 32 79 27 253 809 15.3 5.8 9.3 13.1 26 69 135 197 782 1561 19.2 8.9 12.0 15.1 62 115 179 798 1716 3206 22.5 9.1 13.2 18.4 65 145 267 456 2477 5276

14.5 1.5 5.0 9.4 2 22 69 8 141 449 20.5 5.2 7.9 12.6 22 51 124 101 507 1560 20.8 4.7 9.6 13.1 17 75 135 115 903 2188

21.8 5.3 11.5 17.4 22 107 238 274 1448 4075 23.5 10.1 14.9 20.1 81 179 316 888 3164 5749

1114 A. Moser et al. / Urban Forestry & Urban Greening 14 (2015) 1110–1121

Table 3Results of linear logistic regressions with stem diameter (dbh) as predictor variable and the tree variables tree height, crown base, crown height, crown diameter, cpa,and crown volume as response variables for T. cordata and R. pseudoacacia (equation: ln(response) = a + b × ln(predictor)). The table lists the regression coefficients (a, b),coefficients of determination (r2), residual standard errors (RSE) for bias correction, F-values, p-values, and standard errors (SE) of regression coefficients as well as the samplesize (n).

Species Parameter n a SE b SE r2 RSE F p

Tilia cordata dbh vs tree height 225 0.51 0.10 0.59 0.03 0.79 0.16 863.60 <0.001dbh vs crown base 225 0.35 0.09 0.23 0.03 0.25 0.22 72.41 <0.001dbh vs crown height 225 −0.26 0.09 0.73 0.03 0.76 0.22 724.60 <0.001dbh vs crown diameter 223 −0.61 0.07 0.78 0.02 0.87 0.16 1479.00 <0.001dbh vs cpa 223 −1.47 0.14 1.56 0.04 0.87 0.33 1485.00 <0.001dbh vs crown volume 223 −1.73 0.20 2.29 0.06 0.87 0.48 1504.00 <0.001

Robinia pseudoacacia dbh vs height 195 1.14 0.10 0.42 0.03 0.55 0.22 233.70 <0.001dbh vs crown base 195 0.85 0.16 0.13 0.04 0.04 0.35 8.47 <0.001dbh vs crown height 195 0.38 0.12 0.54 0.03 0.59 0.25 279.70 <0.001dbh vs crown diameter 195 −0.01 0.09 0.60 0.02 0.76 0.19 615.90 <0.001

0.17 1.20 0.05 0.76 0.38 613.80 <0.0010.23 1.74 0.06 0.79 0.50 743.60 <0.001

Cep0sf

dfaThttbtp(ss

tcbcdcttrhrb5etd

wi(sds(ab

Table 4Regression statistics of the form ln(tree height) = a + b × ln(dbh) + CF with dbh aspredictor variable and tree height as response.

Species City a b p

T. cordata München 0.52 0.59 <0.001Würzburg 0.50 0.59 <0.001

dbh vs cpa 195 −0.25

dbh vs crown volume 195 0.13

rown base showed only a very weak association with stem diam-ter for both species. The correlation between dbh and height of R.seudoacacia was highly significant; however, the r2 value was only.55. Comparing both species, T. cordata revealed stronger relation-hips than R. pseudoacacia. Residual standard errors (RSE) rangedrom 0.16 to 0.50, with the maximum value for crown volume.

Fig. 2 illustrates the log–log relationships of tree height, crowniameter, and crown volume with diameter at breast height (dbh)or both species. The models were fitted to the sampled data points,nd the regression equation and associated r2 values are shown.he weaker relationships of tree dimensions (particularly of dbh toeight) of R. pseudoacacia compared to T. cordata possibly reflecthat R. pseudoacacia is a very light-demanding species and thathe measured trees grew unshaded as well as in the shade of highuildings or sheltered by other trees in parks. The growth alloca-ion of trees can greatly change in response to light availability,articularly with respect to the shape of light-demanding speciesHarja et al., 2012; Vincent and Harja, 2008; Weiner, 2004). Thetronger relationship between the height and stem diameter of thehade-tolerant T. cordata supports this idea.

To illustrate the growth of T. cordata and R. pseudoacacia overime, the allometric relationships of height, crown diameter, andrown volume were plotted against age (Fig. 3). The relationshipsetween age and tree dimensions show excellent fits (r2 > 0.8) for T.ordata and moderate (r2 > 0.55) to strong fits (r2 > 0.7) for R. pseu-oacacia. The development of space requirements of both speciesan be estimated from these regressions. The results in Table 3 andhe plots shown in Figs. 2 and 3 present relationships betweenhe age and tree dimensions of both tree species in urban envi-onments. All growth curves show the development from low toigh age and represent a large part of a tree’s lifecycle in urban sur-oundings. Therefore, it is possible to estimate the space occupiedy a tree for any stem diameter and age. For example, at an age of0, T. cordata will have an average height of 18 m, a crown diam-ter of 12.5 m, and a crown volume of 1379 m3. In comparison, athe same age R. pseudoacacia will have a height of 17 m, a crowniameter of 11 m, and a crown volume of 1031 m3.

Our results for the growth patterns of T. cordata show slightlyeaker relationships for tree dimensions than do those of the stud-

es conducted by Larsen and Kristoffersen (2002) and Hasenauer1997). This difference is possibly based on the high variation of ourampled trees, which cover two cities with different climatic con-itions, each with three different types of sites (park, public place,

treet). Compared with the results by Larsen and Kristoffersen2002) who studied T. cordata in Copenhagen, T. cordata in Münchennd Würzburg grew marginally larger. Considering the growth ofoth species in the studied cities in detail, we found only minor

R. pseudoacacia München 1.21 0.40 <0.001Würzburg 1.04 0.45 <0.001

differences for all tree parameters. Regression analysis revealedsmall alterations of the intercept (a), as Table 4 displays exemplar-ily for the tree dimensions stem diameter and height. The slope (b)of T. cordata is similar in München and Würzburg, while R. pseu-doacacia shows a slightly faster height development in Würzburg(b = 0.4 in München to b = 0.45 in Würzburg).

To further interpret the growth patterns of T. cordata and R.pseudoacacia in urban surroundings, growth changes over agewere investigated. The development of both species reveals slightchanges over age; Fig. 4 exemplarily highlights the change in heightin relation to age. While there is a minor difference in the growthpatterns of R. pseudoacacia with age, T. cordata trees show a changedallometry with ages higher than 40 years. This is in line with thefindings by Bertram (1989) and Genet et al. (2011) that constantscaling coefficients can mean an oversimplification especially inadvanced tree development phases. The same effect can be foundwhen regarding the branching allometry of both species. For T.cordata a shift toward less crown diameter increase with increas-ing dbh and age was found, similar to the growth patterns Dahleand Grabosky (2010) described for Acer platanoides. R. pseudoacaciashowed only minor shifts in allometry.

These results support city planners in selecting tree species forpublic places, streets, and parks. Future space requirements of T.cordata and R. pseudoacacia trees in urban environments can becalculated on the basis of the presented relationships, and the func-tions and benefits of trees can be derived. Shading (Akbari et al.,2001), carbon storage (Yoon et al., 2013), cooling (Akbari et al.,2001), fine particle filtering (Bolund and Hunhammar, 1999), noisereduction (Bolund and Hunhammar, 1999; Troxel et al., 2013), andother ecosystem services can be estimated from growth patternsand tree dimensions.

Site differences

Analysis of variance (ANOVA) was used to determine the influ-ence of tree species and site (park, public place, street) on treestructure and growth. Tables 5 and 6 present mean values and

A. Moser et al. / Urban Forestry & Urban Greening 14 (2015) 1110–1121 1115

F 2) croc

si

trffohpbgsadc

ig. 2. Allometric logistic relationships between stem diameter and (1) tree height, (haracteristics of the regression lines are presented in Table 3.

tandard deviations for all measured and calculated tree variablesn the three categories and the ANOVA results.

The variables dbh, age, tree height, crown base, crown diame-er, crown projection area, crown volume, and vitality of T. cordataevealed no significant differences among the site categories. Onlyor crown height and open surface area, park trees considerably dif-ered from street trees (p = 0.02 for crown height and p < 0.001 forpen surface area). T. cordata planted in public places had crowneights similar to those of park and street trees (Table 5). For R.seudoacacia, all tree variables except vitality markedly differedetween site types (p = 0.13 for vitality). Park trees had markedlyreater tree dimensions than street trees. Trees in public places had

tem diameters, ages, tree heights, crown heights, and open surfacereas similar to those of park and street trees (p < 0.001). For crowniameter, crown projection area, and crown volume, R. pseudoaca-ia at public places significantly differed from both park and street

wn diameter, and (3) crown volume for T. cordata and R. pseudoacacia. The statistical

trees (p < 0.001). The overall vitality of T. cordata was slightly betterthan that of R. pseudoacacia, with means of 0.9 and 1.2, respectively,both ranged from very good to good (p = 0.3).

The results presented in Tables 5 and 6 indicate that T. cordatagrew in a similar pattern independent of the site type, whereas R.pseudoacacia was more variable in its growth dimensions based onthe site type. ANOVA revealed that park trees of R. pseudoacacia aremarkedly older than street trees (p < 0.001), an age difference thatshould be carefully considered. Owing to this difference, park treesmay have greater tree dimensions, possibly leading to misinterpre-tations of the impact of site conditions on tree growth.

Some of the species differences in the growth patterns related

to the growth site, city, and age can also be caused by the opensurface area of the tree (tree pit). Regression analyses indicate thatthe growth of T. cordata and R. pseudoacacia is highly influenced bythe tree pit (Table 7). Sanders et al. (2013) and Grabosky and Gilman

1116 A. Moser et al. / Urban Forestry & Urban Greening 14 (2015) 1110–1121

F ight, (a

(psao

sapTghCma

ig. 3. Allometric nonlinear least square relationships between age and (1) tree hend R. pseudoacacia, with resulting r2 values.

2007) as well found significant growth differences in relation tolanting space, revealing that trees with reduced planting spacehow reduced maximum size. Since reduced planting space induces

limited soil volume and limited water availability, tree growth isften inhibited leading to higher tree mortality.

Further, because park trees may be exposed to less watertress, higher growth rates in parks may be expected (Kjelgrennd Clark, 1992; Whitlow and Bassuk, 1988). Our results sup-ort this expectation for most tree dimensions of R. pseudoacacia.hey revealed smaller tree dimensions in streets but considerablyreater dimensions in parks. This difference can reflect light and

ence photosynthetic limitations in street canyons (Kjelgren andlark, 1992). Kjelgren and Clark (1992) investigated the microcli-ates and growth of urban trees in park, streets, and public places

nd found marked growth differences between the sites. Trees in

2) crown diameter, and (3) crown volume of the equation yY= aY× xb for T. cordata

public places had a substantially lower leaf area and diameter incre-ment, a finding mainly explained by higher water stress in plazas.Depending on the species, differences between site conditions canconsiderably influence the growth of urban trees.

Ecosystem services

The statistical relationships between tree dimensions and LAI(m2/m2) derived with generalized additive models are presentedin Table 8.

The mean LAI of T. cordata was 2.60, ranging from 0.98 to 5.29,

while that of R. pseudoacacia trees was 1.49, ranging from 0.42 to5.0. Rauner (1976) found a higher LAI of 4.78 for T. cordata. The LAI ofR. pseudoacacia is lower than that stated by Dickmann et al. (1985)for pure plantings (4.3–4.9) and mixed stands (3.2–3.3) in the US

A. Moser et al. / Urban Forestry & Urban Greening 14 (2015) 1110–1121 1117

Fig. 4. Allometric logistic relationships between stem diameter and tree height for T. cordata and R. pseudoacacia at an age lower than 40 (black) and over 40 (grey). Upperregression equations are for age lower than 40, lower equations are for age over 40.

Table 5Mean values of the measured and calculated tree variables age, dbh, tree height, crown base, crown height, and related standard deviation (SD) in response to growth site forT. cordata and R. pseudoacacia and p-values for each ANOVA. Mean values in the same column differ significantly when followed by different letters (Tukey’s test, p < 0.05),n = sample size.

Site n Age [predicted] DBH [cm] Tree height [m] Crown base [m] Crown height [m]

Mean SD Mean SD Mean SD Mean SD Mean SD

p = 0.16 p = 0.18 p = 0.08 p = 0.13 p = 0.02

Tilia cordataPark 61 44 a 23 30.5 a 15.9 13.0 a 4.4 2.9 a 0.7 10.1 a 4.1Public place 69 48 a 24 33.6 a 16.6 12.8 a 4.2 3.1 a 0.8 9.7 ab 3.9Street 95 41 a 20 28.6 a 14.3 11.6 a 3.9 3.1 a 0.8 8.4 b 3.5

p < 0.001 p < 0.001 p < 0.001 p < 0.001 p < 0.001

Robinia pseudoacaciaPark 45 78 a 39 49.7 a 25.2 18.6 a 5.1 4.9 a 2.7 13.7 a 4.5Public place 63 60 ab 25 38.6 ab 16.5 14.1 b 4.1 3.7 b 1.1 10.4 b 4.4Street 87 55 b 29 35.6 b 18.9 12.9 b 3.7 3.7 b 1.1 9.2 b 3.3

Table 6Mean values of the measured and calculated tree variables crown diameter, crown projection area (cpa), crown volume, open surface area, vitality, and related standarddeviation (SD) in response to growth site for T. cordata and R. pseudoacacia and p-values for each ANOVA. Mean values in the same column differ significantly when followedby different letters (Tukey’s test, p < 0.05), n = sample size.

Site n Crown diameter [m]* CPA [m2]* Crown volume [m3]* Open surface area [m2]* Vitality

Mean SD Mean SD Mean SD Mean SD Mean SD

p = 0.23 p = 0.24 p = 0.11 p < 0.001 p = 0.13

Tilia cordataPark 61 7.9 a 2.9 55 a 42 701 a 889 40.0 a 33.8 0.8 a 0.6Public place 69 8.2 a 3.3 61 a 43 723 a 691 26.1 b 26.9 1.0 a 0.6Street 95 7.3 a 3.0 48 a 38 525 a 595 19.5 b 21.7 0.8 a 0.6

p < 0.001 p < 0.001 p < 0.001 p < 0.001 p = 0.35

Robinia pseudoacaciaPark 45 10.9 a 3.9 105 a 72 1589 a 1362 74.7 a 59.1 1.3 a 0.8Public place 63 8.9 b 2.6 67 b 37 805 b 687 17.7 b 24.2 1.3 a 0.7

36

dL

ob

dta

Street 87 7.9 c 2.7 55 c

* n = 67 of T. cordata at public places.

uring the early growing season but very similar to the computedAI in August (1.7).

Based on the calculated LAI and biomass, the ecosystem servicesf shading, carbon storage, and cooling effect were estimated foroth species of different age classes (Table 9).

The carbon storage capacity of both tree species considerablyiffers (t = −8.12, p < 0.001), with R. pseudoacacia storing more thanwice as much carbon than T. cordata in average. Although the aver-ge carbon storage of a T. cordata tree at the age of 80 years can reach

596 c 657 12.0 b 21.4 1.1 a 0.6

1341 kg, R. pseudoacacia of the same age can store up to 2119 kg C.A maximum storage capacity of more than 3000 kg C was found forpark and street trees of R. pseudoacacia. These results are similarto those by Yoon et al. (2013), who developed growth equationsto quantify the aboveground biomass and carbon storage of five

urban tree species in Daegu, Korea. They found a carbon storagecapacity of Platanus orientalis ranging from 100 kg C in young treesto 350 kg C at a stem diameter of 50 cm. R. pseudoacacia achievedhigher benefits than T. cordata up to an age of 40 to 60 years

1118 A. Moser et al. / Urban Forestry & Urban Greening 14 (2015) 1110–1121

Table 7Results of linear logistic regressions with tree pit (open surface area) as predictor variable and the tree variables dbh, tree height, crown base, crown diameter, crown height,cpa, crown volume, and vitality as response variables for T. cordata and R. pseudoacacia (equation: ln(response) = a + b × ln(predictor)). The table lists the regression coefficients(a, b), coefficients of determination (r2), residual standard errors (RSE) for bias correction, F-values, p-values, and standard errors (SE) of regression coefficients as well as thesample size (n).

Species Parameter n a SE b SE r2 RSE F p

Tilia cordata Tree pit vs dbh 221 2.36 0.09 0.33 0.03 0.35 0.44 122.2 <0.001Tree pit vs height 221 1.84 0.06 0.21 0.02 0.35 0.29 118.3 <0.001Tree pit vs crown base 221 1.00 0.05 0.03 0.02 0.01 0.25 3.9 0.05Tree pit vs crown diameter 221 1.12 0.07 0.30 0.02 0.41 0.35 157.7 <0.001Tree pit vs crown height 221 1.34 0.07 0.28 0.02 0.37 0.36 133.1 <0.001Tree pit vs cpa 221 1.99 0.14 0.59 0.05 0.41 0.70 158.1 <0.001Tree pit vs crown volume 221 3.33 0.21 0.87 0.07 0.42 1.02 161.1 <0.001Tree pit vs vitality 221 1.21 0.12 −0.12 0.04 0.03 0.59 8.6 0.004

Robinia pseudoacacia Tree pit vs dbh 193 2.97 0.08 0.22 0.03 0.23 0.50 58.4 <0.001Tree pit vs height 193 2.27 0.04 0.14 0.02 0.29 0.27 81.8 <0.001Tree pit vs crown base 195 1.21 0.06 0.04 0.02 0.01 0.36 3.7 0.06Tree pit vs crown diameter 193 1.70 0.05 0.16 0.02 0.27 0.33 72.7 <0.001Tree pit vs crown height 193 1.85 0.06 0.17 0.02 0.28 0.34 76.7 <0.001Tree pit vs cpa 195 3.15 0.11 0.32 0.04 0.27 0.67 72.6 <0.001Tree pit vs crown volume 193 5.00 0.15 0.49 0.05 0.31 0.92 88.1 <0.001Tree pit vs vitality 193 1.35 0.11 −0.05 0.04 0.01 0.68 2.0 0.16

Table 8Model parameters and sample size (n) of the calculated models for T. cordata [ln(LAI) = a + s(b1) × ln(dbh) + s (b2) × ln(tree height) + s(b3) × ln(crown diameter)] and R.pseudoacacia [ln(LAI) = a + s (b1) × ln(tree height) + s (b2) × ln(crown volume) + s (b3) × ln(CPA)] with smooth function s for every slope variable (b).

Species n a SE r2 p

Tilia cordata 193 0.93 0.03 0.35 <0.001Robinia pseudoacacia 129 0.26 0.07 0.22 <0.001

Table 9Mean, minimum (min), and maximum (max) predicted carbon storage in kg C per tree, and shaded area in m2 per tree, as well as the cooling effect as mean energy removedfrom the atmosphere per tree in kW h in summer month for T. cordata and R. pseudoacacia from planting at <20 to high age >80 in ascending age classes with overall averagevalues for all trees per species and sample size (n).

Age n C-storage [kg C] Shading [m2] Energy removed fromatmosphere [kW h]

Min Mean Max Min Mean Max

Tilia cordata<20 27 3 8 17 18 36 75 51720–40 86 17 62 126 32 144 323 255340–60 83 124 219 356 131 310 505 635060–80 22 376 494 714 327 533 843 10,797>80 8 781 1341 2571 192 718 1171 14,772

Average 196 246 4843Robinia pseudoacacia

<20 34 11 55 132 19 96 275 116420–40 67 99 207 402 74 230 604 291640–60 47 124 424 856 93 341 701 4797

rrvectcapb

ica

aa

60–80 33 289 746 1990

>80 14 718 2119 3284

Average 461

egarding the ecosystem functions of shading and cooling (energyemoved from the atmosphere). With higher age, T. cordata pro-ided more cooling and similar shade than R. pseudoacacia. Rahmant al. (2011) analyzed the cooling ability by transpiration of Pyrusalleryana and found maximum values of 7500 W for ten year oldrees in August (max. 180 kW h per tree in one month), which areomparable values to the young T. cordata trees in our study. Inverage, R. pseudoacacia can provide significant higher shade com-ared to T. cordata (t = −3.95, p < 0.001), while the cooling ability ofoth species is more similar (t = 0.39, p = 0.69).

Fig. 5 reveals the average carbon storage, shading, and cool-ng effect of T. cordata and R. pseudoacacia for the different siteategories park, public place, and street; Table 10 provides the

ssociated p-values for each comparison.

Street trees of R. pseudoacacia provided the least carbon stor-ge (F = 9.9, p < 0.001), shading (F = 16.8, p < 0.001), and coolingbility (F = 17.4, p < 0.001) of the measured sites, while park trees

201 456 913 6683334 781 1188 13,644

312 4472

exhibited the highest ecosystem services compared to public placesand street. T. cordata trees in public places had higher ecosystemservices than did park and street trees, but all values were not sig-nificant (carbon storage: F = 1.8, p = 0.16, shading: F = 2.6, p = 0.07,cooling effect: F = 2.3, p = 0.10). Because park trees of R. pseudoaca-cia were markedly older than trees at streets and public places, thegreater services also may be because of the age differences.

Urban trees can considerably improve city climates by provid-ing ecosystem services. The results of our research support thoseof other studies dealing with microclimates amelioration by tem-perature reduction (Souch and Souch, 1993), shading and cooling(Akbari et al., 2001; Dimoudi and Nikolopoulou, 2003), humidityincrease (Dimoudi and Nikolopoulou, 2003; Georgi and Zafiriadis,

2006), and air quality improvement (Akbari et al., 2001). Furtherinvestigations will enable comparisons of tree species growth andtheir benefits to urban green space planning. Particularly withdendrochronology analyses and the integration of climate and

A. Moser et al. / Urban Forestry & Urban Greening 14 (2015) 1110–1121 1119

Fig. 5. Box plots of the ecosystem services: carbon storage, shading, and cooling effect of T. cordata (white) and R. pseudoacacia (gray) and for different sites (park, publicplace, and street).

Table 10p-Values of conducted analysis of variance (ANOVA, Tukey’s test) for differences regarding the ecosystem services of T. cordata (T, right upper part) and R. pseudoacacia (R,left lower part) for growing sites park (Pa), public place (Pl), and street (St), significant differences marked bold.

Carbon storage Shading Cooling effect

R\T Pa Pl St R\T Pa Pl St R\T Pa Pl St

3

1

soMtbmoamdpl

C

ttgttidoalBl2f

Pa 0.88 0.43 Pa

Pl 0.01 0.16 Pl 0.00St <0.001 0.27 St <0.00

oil conditions in tree growth modeling will improve knowledgef urban tree growth patterns in association with their services.oreover, future studies should be based on specific biomass func-

ions for T. cordata and R. pseudoacacia, such as those generatedy terrestrial laser scanning. With scanned images of trees, infor-ation about leaf distribution and crown transparency can be

btained. The shaded area was calculated based on crown heightnd diameter, while crown transparency was not considered. Withore precise information about leaf area, leaf density, and leaf

istribution within the tree crown, accurate shading values androcess-based modeling of urban tree growth can be derived from

aser scanning and improved, respectively.

onclusions

For an adequate planning of urban green areas and the realiza-ion of certain demands (esthetics and spatial functions), modelinghe growth patterns of urban trees is crucial. Because empiricalrowth modeling of urban trees is scarce, we analyzed the growth ofhe common urban tree species T. cordata and R. pseudoacacia. Withhe shown conversion equations, dbh at 1.3 m could be transformednto stem diameter at 1 m. By using the identified relationships treeimensions of T. cordata and R. pseudoacacia in urban environmentsf south German cities could be predicted for any stem diameternd age. The growth patterns of both species were strongly corre-ated; though R. pseudoacacia showed slightly weaker relationships.

ecause urban trees face challenges, such as less water availability,

imited root volume (Grabosky and Gilman, 2007; Sanders et al.,013), and higher temperatures (Akbari et al., 2001), tree growthor the site categories park, public place, and street canyon were

0.80 0.08 Pa 0.97 0.240.28 Pl 0.001 0.14

0.04 St <0.001 0.04

analyzed. The results revealed similar growth patterns of T. cordata,regardless of their site while R. pseudoacacia exhibited markedlybetter growth in parks. Other factors like age and open surface areahad influence on the growth patterns of both species. Regressionanalysis confirmed the importance of the open surface area for thedevelopment of urban trees. The estimation of ecosystem serviceslike carbon storage, shading, and cooling by transpiration indicatethat T. cordata and R. pseudoacacia can improve the urban micro-climate. Thus, the ecosystem services provided by both species arebeneficial for the city climate and increase human thermal comfort(Bolund and Hunhammar, 1999; Georgi and Zafiriadis, 2006). Anunderstanding of the development of single tree dimensions canfacilitate benefit assessment that trees provide for urban climates.The results indicate the virtue of urban trees, however some areup to now only rough estimates. The present study facilitates treeselection based on space requirements and their functions for thecity climate. Up to now, only limited data are available for urbantree growth in central Europe; more research dealing with allom-etry and functions of trees in urban areas is required.

Acknowledgments

This research was funded by the Bavarian State Ministry of theEnvironment and Consumer Protection (project TUF01UF-64971“Urban trees under climate change: their growth, environmentalperformance, and perspectives”). We thank the departments for

the municipal green areas of München and Würzburg for theirsupport and encouragement. The authors also thank Chan Ka Nok,Alexander Hellwig, and Karin Beer for their assistance in field datacollection.

1 Urban

R

A

B

B

B

B

B

B

C

C

C

C

D

D

D

D

D

D

DD

F

G

G

G

G

HH

H

I

K

K

L

L

M

M

M

120 A. Moser et al. / Urban Forestry &

eferences

kbari, H., Pomerantz, M., Taha, H., 2001. Cool surfaces and shade trees to reduceenergy use and improve air quality in urban areas. Solar Energy 70, 295310.

askerville, G.L., 1971. Use of logarithmic regression in the estimation of plantbiomass. Can. J. For. 2, 49–53.

eatty, R.A., Heckman, C.T., 1981. Survey of urban tree programs in the UnitedStates. Urban Ecol. 5, 81–102.

ertram, J.E.A., 1989. Size-dependent differential scaling in branches: themechanical design of trees revisited. Trees 3, 241–253.

inkley, D., Campoe, O.C., Gspaltl, M., Forrester, D.I., 2013. Light absorption and useefficiency in forests: why patterns differ for trees and stands. For. Ecol.Manage. 288, 5–13.

olund, P., Hunhammar, S., 1999. Ecosystem services in urban areas. Ecol. Econ. 29,293–301.

onhomme, R., Chartier, P., 1972. The interpretation and automatic measurementof hemispherical photographs to obtain sunlit foliage area and gap frequency.Israel J. Agric. Res. 22, 53–61.

hiba, Y., 1998. Architectural analysis of relationship between biomass and basalarea based on pipe model theory. Ecol. Model. 108, 219–225.

hreptun, C., 2015. Kronenstruktureigenschaften von Linden und Robinien inMünchen. Anwendungen des terrestrischen Laserscannings. Chair for ForestGrowth and Yield Science. Technische Universität München, Freising, M. Sc.Thesis.

ˇ ihák, T., Hlásny, T., Stolariková, R., Vejpustková, M., Marusák, R., 2014. Functionsfor the aboveground woody biomass in small-leaved lime (Tilia cordata Mill.).For. J. 60, 150.

lark III, A., Schroeder, J.G., 1986. Weight, Volume and Physical Properties of MajorHardwood Species in the Southern Appalachian Mountains. SoutheasternForest Experiment Station, United States Department of Agriculture, ForestService.

ahle, G.A., Grabosky, J.C., 2009. Review of literature on the function and allometricrelationships of tree stems and branches. Arboricult. Urban For. 35, 311–320.

ahle, G.A., Grabosky, J.C., 2010. Allometric patterns in Acer platanoides(Aceraceae) branches. Trees 24, 321–326.

ay, S.D., Bassuk, N.L., van Es, H., 1995. Effects of four compaction remediationmethods for landscape trees on soil aeration, mechanical impedance and treeestablishment. J. Arboricult. 22, 144–150.

ickmann, D.I., Steinbeck, K., Skinner, T., 1985. Leaf area and biomass in mixed andpure plantations of sycamore and black locust in the Georgia Piedmont. For.Sci. 31, 509–517.

imoudi, A., Nikolopoulou, M., 2003. Vegetation in the urban environment:microclimatic analysis and benefits. Energy Build. 35, 69–76.

VWK, 1996. Ermittlung der Verdunstung von Land- und Wasserflächen. In:DVWK-Merkblätter zur Wasserwirtschaft. Wirtschafts- und Verl.-Ges. Gas undWasser. DVWK, Bonn.

WD, 2015. Deutscher Wetterdienst.wyer, J.F., 2009. How old is that tree? Illinois Trees: The quarterly publication of

the Illinois Arborist Association 24, 13.oster, R.S., Blaine, J., 1978. Urban trees survival: trees in the sidewalk. J.

Arboricult. 4, 14–17.enet, A., Wernsdörfer, H., Jonard, M., Pretzsch, H., Rauch, M., Ponette, Q., Nys, C.,

Legout, A., Ranger, J., Vallet, P., Saint-André, L., 2011. Ontogeny partly explainsthe apparent heterogeneity of published biomass equations for Fagus sylvaticain central Europe. For. Ecol. Manage. 261, 1188–1202.

eorgi, N.J., Zafiriadis, K., 2006. The impact of park trees on microclimate in urbanareas. Urban Ecosyst. 9, 195–209.

rabosky, J., Bassuk, N.L., 1995. A new urban tree soil to safely increase rootingvolumes under sidewalks. J. Arboricult. 21, 187–201.

rabosky, J.C., Gilman, E.F., 2007. Response of two oak species to reduction pruningcuts. Arboricult. Urban For. 33, 360–366.

äckel, H., 2012. Meteorologie. UTB, Eugen Ulmer, Stuttgart.arja, D., Vincent, G., Mulia, R., van Noordwijk, M., 2012. Tree shape plasticity in

relation to crown exposure. Trees 26, 1275–1285.asenauer, H., 1997. Dimensional relationship of open-grown trees in Austria. For.

Ecol. Manage. 96, 197–206.PCC, 2003. The Intergovernmental Panel on Climate Change. Good Practice

Guidance for Land Use, Land-Use Change and Forestry. Institute for GlobalEnvironmental Strategies for the IPCC, Kanagawa.

jelgren, R.K., Clark, J.R., 1992. Microclimates and tree growth in tree urban spaces.J. Environ. Hortic. 10, 139–145.

ramer, J.K., Kozlowski, T.T., 1979. Physiology of Woody Plants. Academic Press,London.

arsen, F.K., Kristoffersen, P., 2002. Tilia’s physical dimensions over time. J.Arboricult. 28, 209–214.

ukaszkiewicz, J., Kosmala, M., 2008. Determining the age of streetside trees withdiameter at breast height-based multifactorial model. Arboricult. Urban For.34, 137–143.

äkelä, A., 1990. Modelling structural–functional relationships in whole-treegrowth; resource allocation. In: Dixon, R.K., Meldahl, R.S., Ruark, G.A., Warren,W.G. (Eds.), Process Modelling of Forest Growth Responses to Environmental

Stress. Timber Press, Portland, OR, pp. 81–95.

cMahon, T.A., 1975. Using body size to understand the structural design ofanimals: quadrupedal locomotion. J. Appl. Physiol. 39, 619–627.

cPherson, E.G., 1998. Atmospheric carbon dioxide reduction by Sacramento’surban forest. J. Arboricult. 24, 215–223.

Greening 14 (2015) 1110–1121

McPherson, E.G., Nowak, D.J., Heisler, G., Grimmond, S., Souch, C., Grant, R.,Rowntree, R.A., 1997. Quantifying urban forest structure, function, and value:the Chicago Urban Forest Climate Project. Urban Ecosyst. 1, 49–61.

McPherson, E.G., Peper, P.J., 2012. Urban tree growth modeling. Arboricult. UrbanFor. 38, 172–180.

McPherson, E.G., Simpson, J.R., Peper, P.K., Scott, K.I., Xiao, Q., 2000. Tree Guidelinesfor Coastal Southern California Communites. In: Local GovernmentCommission and Western Center for Urban Forest Research and Education,Sacramento, CA.

Miller, J.B., 1967. A formula for average foliage density. Aust. J. Bot. 15, 141–144.Niklas, K.J., 1994. Plant Allometry. The scaling of form and process. University of

Chicago Press, Chicago, IL.Nock, C.A., Caspersen, J.P., Thomas, S.C., 2008. Large ontogenetic declines in

intra-crown leaf area index in two temperate deciduous tree species. Ecology89, 744–753.

Nowak, D.J., Crane, D.E., 2002. Carbon storage and sequestration by urban trees inthe USA. Environ. Pollut. 116, 381–389.

Nowak, D.J., Kuroda, M., Crane, D.E., 2004. Tree mortality rates and tree populationprojections in Baltimore, Maryland, USA. Urban For. Urban Green. 2, 139–147.

Nowak, D.J., McBride, J., Beatty, R., 1990. Newly planted street tree growth andmortality. J. Arboricult. 27, 306–317.

Pauleit, S., Jones, N., Garcia-Marin, G., Garcia-Valdecantos, J.-L., Rivière, L.M.,Vidal-Beaudet, L., Bodson, M., Randrup, T.B., 2002. Tree establishment practicein towns and cities—results from a European survey. Urban For. Urban Green.1, 83–96.

Peper, P.J., Alzate, C.P., McNeil, J.W., Hashemi, J., 2014. Allometric equations forurban ash trees (Fraxinus spp.) in Oakville, Southern Ontario, Canada. UrbanFor. Urban Green. 13, 175–183.

Peper, P.J., McPherson, E.G., 1998. Comparison of five methods for estimating leafarea index of open grown deciduous trees. J. Arboricult. 24, 98–111.

Peper, P.J., McPherson, E.G., Mori, S., 2001a. Equations for predicting diameter,height, crown width, and leaf area of San Joaquin Valley street trees. J.Arboricult. 27, 306–317.

Peper, P.J., McPherson, E.G., Mori, S.M., 2001b. Predictive equations for dimensionsand leaf area of coastal Southern California street trees. J. Arboricult. 27,169–180.

Pretzsch, H., 2014. Canopy space filling and tree crown morphology inmixed-species stands compared with monocultures. For. Ecol. Manage. 327,251–264.

Pretzsch, H., Biber, P., Uhl, E., Dahlhausen, J., Rötzer, T., Caldentey, J., Koike, T., vanCon, T., Chavanne, A., Seifert, T., du Toit, B., Farnden, C., Pauleit, S., 2015. Crownsize and growing space requirement of common tree species in urban centres,parks, and forests. Urban For. Urban Green. 14, 466–479.

Pretzsch, H., Matthew, C., Dieler, J., 2012. Allometry of tree crown structure.Relevance for space occupation at the individual plant level and forself-thinning at the stand level. In: Matyssek, R.e.a. (Ed.), Growth and Defencein Plants. Springer Verlag, Berlin Heidelberg.

Quigley, M.F., 2004. Street trees and rural conspecifics: will long-lived trees reachfull size in urban conditions? Urban Ecosyst. 7, 29–39.

R Core Team, 2014. R: A Language and Environment For Statistical Computing. RFoundation for Statistical Computing, Vienna, Austria.

Rahman, M.A., Smith, J.G., Stringer, P., Ennos, A.R., 2011. Effect of rootingconditions on the growth and cooling ability of Pyrus calleryana. Urban For.Urban Green. 10, 185–192.

Rauner, J.L., 1976. Deciduous forests. In: Monteith, J.L. (Ed.), Vegetation and theAtmosphere 2: Case Studies. Academic Press, London, p. 241.

Rhoades, R.W., Stipes, R.J., 1999. Growth of trees on Virginia Tech campus inresponse to various factors. J. Arboricult. 25, 211–217.

Richards, N.A., 1979. Modeling survival and consequent replacement needs of astreet tree population. J. Arboricult. 5, 251–255.

Rijal, B., Weiskittel, A.R., Kershaw, J.A., 2012. Development of height to crown basemodels for thirteen tree species of the North American Acadian Region. For.Chronicle 88,60–73.

Roloff, A., 2001. Baumkronen. In: Verständnis und praktische Bedeutung eineskomplexen Naturphänomens. Ulmer, Stuttgart.

Roman, L.A., Scatena, F.N., 2011. Street tree survival rates: meta-analysis ofprevious studies and application to a field survey in Philadelphia, PA, USA.Urban For. Urban Green. 10, 269–274.

Rötzer, T., Leuchner, M., Nunn, A.J., 2010. Simulating stand climate, phenology, andphotosynthesis of a forest stand with a process-based growth model. Int. J.Biometeorol. 54, 449–464.

Rust, S., 2014. Analysis of regional variation of height growth and slenderness inpopulations of six urban tree species using a quantile regression approach.Urban For. Urban Green. 13, 336–343.

Ryan, M.G., Binkley, D., Fownes, J.H., 1997. Age-related decline in forestproductivity: pattern and process. In: Begon, M., Fitter, A.H. (Eds.), Advances inEcological Research. Academic Press, San Diego, London, pp. 227–228.

Sanders, J., Grabosky, J., Cowie, P., 2013. Establishing maximum size expectationsfor urban trees with regard to designed space. Arboricult. Urban For. 39, 68–73.

Scott, K.I., McPherson, E.G., Simpson, J.R., 1998. Air pollutant uptake by Sacrameto’s

urban forest. J. Arboricult. 24, 201–214.

Shashua-Bar, L., Hoffman, M.E., 2003. Geometry and orientation aspects in passivecooling of canyon streets with trees. Energy Build. 35, 61–68.

Shinozaki, K.K., Nozumi, Y.K., Kira, T., 1964a. A quantitative analysis of plantform-the pipe model theory. I. Basic analysis. Jpn. J. Ecol. 14, 97–105.

Urban

S

SS

S

S

S

S

T

V

Hydrol. Process. 14, 763–784.

A. Moser et al. / Urban Forestry &

hinozaki, K.K., Yoda, K., Hozumi, K., Kira, T., 1964b. A quantitative analysis of plantform-Pipe model theory. II. Further evidences of the theory and its applicationin forest ecology. Jpn. J. Ecol. 14, 133–139.

kiera, B., Moll, G., 1992. The sad state of city trees. Am. For. 98 (3/4), 61–64.ouch, C.A., Souch, C., 1993. The effect of trees on summertime below canopy

urban climates: a case study Bloomington, Indiana. J. Arboricult. 19, 303–312.prugel, D.G., 1983. Correcting for bias in log-transformed allometric equations.

Ecology 64, 209–210.tage, A.R., 1973. Prognosis model for stand development. In: Research Paper

INT-137. USDA Forest Service, Intermountain Forest and Range ExperimentStation, Ogden, UT (32 pp.).

tevens, J., 1999. Intermediate Statistics. A Modern Approach. Lawrence ErlbaumAssociates, London.

toffberg, G.H., van Rooyen, M.W., van der Linde, M.J., Groeneveld, H.T., 2008.Predicting the growth in tree height and crown size of three street tree speciesin the City of Tshwane, South Africa. Urban For. Urban Green. 7, 259–264.

roxel, B., Piana, M., Ashton, M.S., Murphy-Dunning, C., 2013. Relationshipsbetween bole and crown size for young urban trees in the northeastern USA.Urban For. Urban Green. 12, 144–153.

incent, G., Harja, D., 2008. Exploring ecological significance of tree crownplasticity through three-dimensional modelling. Ann. Bot. 101, 1221–1231.

Greening 14 (2015) 1110–1121 1121

Watt, M.S., Kirschbaum, M.U.F., 2011. Moving beyond simple linear allometricrelationships between tree height and diameter. Ecol. Model. 222,3910–3916.

Weiner, J., 2004. Allocation, plasticity and allometry in plants. Perspect. Plant Ecol.Evol. Syst. 6, 207–215.

Welles, J.M., Norman, J.M., 1991. Instrument for indirect measurement of canopyarchitecture. Agron. J. 83, 818–825.

Whitlow, T.H., Bassuk, N.L., 1986. Trees in difficult sites. J. Arboricult. 13, 10–17.Whitlow, T.H., Bassuk, N.L., 1988. Ecophysiology of urban trees and their

management—The North American experience. HortScience 23, 542–546.Xiao, Q., McPherson, E.G., Ustin, S.L., Grismer, M.E., Simpson, J.R., 2000a. A new

approach to modeling tree rainfall interception. J. Geophys. Res. 105,29–173.

Xiao, Q., McPherson, E.G., Ustin, S.L., Grismer, M.E., Simpson, J.R., 2000b. Winterrainfall interception by two mature open-grown trees in Davis, California.

Yoon, T.K., Park, C.-W., Lee, S.J., Ko, S., Kim, K.N., Son, Y., Lee, K.-H., Oh, S., Lee, W.-K.,Son, Y., 2013. Allometric equations for estimating the aboveground volume offive common urban street tree species in Daegu, Korea. Urban For. UrbanGreen. 12, 344–349.